Method and System for Image Processing to Determine a Region of Interest

1. A method for image processing, said method comprising: in a computing device: determining luminance values of a plurality of pixels in a subset of a frame of a two-dimensional image; determining texture values of said plurality of pixels in said subset of said frame; and identifying a subject region and a background region in said frame of said two-dimensional image based on said determined luminance values and said determined texture values of said plurality of pixels, wherein said identification of said subject region and said background region in said frame of said two-dimensional image comprises classifying said plurality of pixels into a first class and/or a second class based on a comparison of said luminance values and said texture values of said plurality of pixels with a first set of lookup tables and a second set of lookup tables respectively.

2. The method of claim 1 , comprising storing said luminance values and said texture values of said plurality of pixels in said first set of lookup tables and said second set of lookup tables respectively.

3. The method of claim 1 , wherein said first class comprises a head class and said second class comprises a background class, wherein said head class comprises a face class and a hair class.

4. The method of claim 1 , comprising converting said two-dimensional image to a three-dimensional image based on said identification of said subject region and said background region in said two-dimensional image.

5. A method for image processing, said method comprising: in a computing device: generating a first set of values based on luminance of a first class of pixels and a second class of pixels in a first two-dimensional image; and generating a second set of values based on texture of said first class of pixels and said second class of pixels in said first two-dimensional image, wherein said generated first set of values and said generated second set of values are used to classify a subject pixel into one of a first class of pixels and a second class of pixels based on a comparison of a luminance value of said subject pixel and a texture value of said subject pixel with said generated first set of values and said generated second set of values respectively.

6. The method of claim 5 , wherein said subject pixel belongs to said first two dimensional image.

7. The method of claim 5 , wherein said subject pixel belongs to a second two dimensional image different from said first two-dimensional image.

8. The method of claim 5 , wherein said first set of values corresponds to a first set of lookup tables that store histogram counts of luminance values of said first class of pixels and said second class of pixels in said first two-dimensional image.

9. The method of claim 5 , wherein said second set of values corresponds to a second set of lookup tables that store gradient distinctiveness indicator functions for said first class of pixels.

10. A method for image processing, said method comprising: in a computing device: determining a region of interest in a two-dimensional image, comprising: determining a first class of pixels and a second class of pixels in said two dimensional image based on a first set of predetermined values corresponding to a luminance of said two-dimensional image, wherein said determination of said first class of pixels and said second class of pixels in said two-dimensional image comprises: predicting a first region and a second region in said two-dimensional image corresponding to said first class of pixels and said second class of pixels; and computing class-conditional probability functions based on luminance of said first and second class of pixels and said first set of predetermined values; and analyzing texture of said first class of pixels and said second class of pixels based on a second set of predetermined values corresponding to said texture of said two-dimensional image; and converting said two-dimensional image to a three-dimensional image based on said determination of said region of interest in said two-dimensional image.

11. The method of claim 10 , further comprising generating said two-dimensional image from a two-dimensional input image, wherein said two-dimensional input image corresponds to an image frame from a sequence of moving image frames.

12. The method of claim 11 , wherein said generation of said two-dimensional image from said two-dimensional input image comprises cropping, resizing and/or down- quantizing said two-dimensional input image.

13. The method of claim 10 , wherein said class-conditional probability functions are computed using Bayes rule.

14. The method of claim 10 , wherein said predicted first region and said predicted second region comprise a head region and a background region respectively.

15. The method of claim 10 , wherein said first set of predetermined values corresponds to a first set of lookup tables to store histogram counts of luminance values of said first class of pixels and said second class of pixels of one or more of said two-dimensional image and at least one previous two-dimensional image.

16. The method of claim 10 , wherein said analyses of said texture comprises: computing gradient distinctiveness indicator functions for said first class of pixels and said second class of pixels; and determining uniqueness of said first class of pixels as compared to said second class of pixels based on said computed gradient distinctiveness indicator functions and said second set of predetermined values.

17. The method of claim 10 , wherein said second set of predetermined values corresponds to a second set of lookup tables to store said gradient distinctiveness indicator functions for said first class of pixels of one or more of said two-dimensional image and at least one previous two-dimensional image.

18. The method of claim 10 , wherein said first class comprises a head class and said second class comprises a background class, wherein said head class comprises a face class and a hair class.

19. The method of claim 10 , wherein said determined region of interest comprises one or both: of a hair region and/or a head region.

20. The method of claim 10 , wherein said determined region of interest is represented as a Boolean-type image.

21. The method of claim 10 , further comprising: smoothing said determined region of interest; eliminating outliers around said smoothed region of interest; detecting an outer contour of said region of interest to generate a region of interest mask image; and restoring said region of interest mask image to an original scale of a two-dimensional input image.

22. A system for image processing, said system comprising: one or more processors in said computing device being operable to: determine luminance values of a plurality of pixels in a subset of a frame of a two-dimensional image; determine texture values of said plurality of pixels in said subset of said frame based on distinctive gradients of said plurality of pixels; and identify a subject region and a background region in said frame of said two-dimensional image based on said determined luminance values and said determined texture values of said plurality of pixels, wherein said one or more processors are operable to identify said subject region and said background region in said frame of said two-dimensional image by classifying said plurality of pixels into a first class and/or a second class based on a comparison of said luminance values and said texture values of said plurality of pixels with a first set of lookup tables and a second set of lookup tables respectively.

23. The system of claim 22 , further comprising a memory operable to store: said first set of lookup tables corresponding to said determined luminance values; and said second set of lookup tables corresponding to said determined texture values.

24. The system of claim 22 , wherein said first class comprises a head class and said second class comprises a background class, wherein said head class comprises a face class and a hair class.

25. The system of claim 22 , wherein said one or more processors are operable to convert said two-dimensional image to a three-dimensional image based on said identification of said subject region and said background region in said two-dimensional image.

26. A non-transitory computer-readable storage medium having stored thereon, a computer program having at least one code section for image processing, the at least one code section being executable by a computer for causing said computer to perform steps comprising: determining luminance values of a plurality of pixels in a subset of a frame of a two-dimensional image; determining texture values of said plurality of pixels in said subset of said frame; and identifying a subject region and a background region in said frame of said two-dimensional image based on said determined luminance values and said determined texture values of said plurality of pixels, wherein said identification of said subject region and said background region in said frame of said two-dimensional image comprises classifying said plurality of pixels into a first class and/or a second class based on a comparison of said luminance values and said texture values of said plurality of pixels with a first set of lookup tables and a second set of lookup tables respectively.